一种基于参数自整定的小波去噪算法研究  

Research on an Improved Wavelet Denoising Algorithm with Parameter Self-tuning

作　　者：刘宝[1] 冯巨龙 宋树鹏 叶会会 LIU Bao;FENG Ju-long;SONG Shu-peng;YE Hui-hui(College of Information and Engineering,China University of Petroleum,Qingdao 266580,China;China Automotive Technology&Research Center,Tianjin 300300,China)

机构地区：[1]中国石油大学(华东)信息与控制工程学院,山东青岛266580 [2]中国汽车技术研究中心,天津300300

出　　处：《控制工程》2020年第3期444-450,共7页Control Engineering of China

基　　金：国家自然科学基金重点项目(60534020,60775052)。

摘　　要：小波阈值去噪法因其算法简单、计算量小而广泛应用于信号去噪,提高了数据处理的精度和效率。为解决传统小波阈值去噪法中阈值函数连续性差或原始小波系数与估计的小波系数存在恒定偏差的问题,提出了一种基于参数自整定阈值函数的改进小波去噪算法。该方法采用指数平滑逼近对小波系数进行阈值处理,通过参数的自整定使其阈值函数能够很快逼近硬阈值函数曲线,并具有较好的连续性和更佳的阈值估计小波系数,从而使改进后的小波去噪算法具有更好的去噪效果。仿真实验结果表明:相比于传统的硬、软阈值去噪方法,改进后的小波去噪算法的去噪效果更优。Wavelet threshold denoising method is widely used in signal denoising because of its simple algorithm and less calculation,which improves the accuracy and efficiency of data processing.To solve the problem of poor continuity of threshold function and constant deviation between the original values and estimated value of wavelet coefficients in traditional wavelet threshold denoising method,an improved wavelet denoising algorithm based on parameter self-tuning threshold function is proposed.By adopting exponential smoothing approach concept and the self-tuning of the parameter,the threshold function can rapidly approach the curve of hard threshold function,and has better continuity and better estimated wavelet coefficients.The simulation results show that the improved wavelet threshold denoising algorithm has better filtering effect compared with the traditional hard and soft threshold denoising methods.

关 键 词：小波去噪 参数自整定 阈值函数 

分 类 号：TN911.4[电子电信—通信与信息系统]